Two ad execs have launched an AI-driven platform that helps advertisers bid efficiently on Google’s ad channels, especially digital and e-commerce.

Amanda AI, which has been operating in Europe for three years, helps businesses manage the long tail of search and shopping keywords. 

Hertz, Xerox, Logitech have used the platform, which has delivered up to five-fold increase in RoI for brands.

SEM: Automatic for the people

Created in Sweden, Amanda AI enables companies to optimise their Google search and shopping campaigns by applying machine learning software to campaign management.

Now launching in Australia via two Sydney-based admen Ben Taylor (ex-Vizuem) and Jonas Lembke (ex-BBDO), the business works by undertaking an exploratory audit for any company that runs search marketing and enables them to access the platform via a self-service model. 

Amanda AI focuses on working with retailers and other businesses with large product sets that need to effectively manage the long tail of their product catalogue.

The platform automatically indexes customers' pages and picks up information about URLs, products and words, and enters the information into a data model.

That data is fed into the Google keyword bidding program and the platform can then optimise paid media campaigns. 

Long tail, short work

“Humans aren’t equipped for long tail optimisations,” said Taylor, Managing Director of technology design studio The Innovation Chapel, which has partnered with Amanda AI to bring the company to Australia.

“We are fixing the problem of managing the Google marketing platform by providing an AI-based approach to keyword optimisation and search marketing.’’

He added: “If you are a fashion retailer with ten brands with twenty different products in three different colours and three different sizes, then that is over 140,400,000 possible bids you will need to make in broad, phrase and exact keyword matching across all demographics and other variables.” 

Amanda AI calculates, based on the probability of a consumer undertaking a transaction online, what price a company should pay for the click, defines its target audience, and adjusts bid prices for future ad spend. 

“Google ads used to be a channel surrounded by a lot of administration. It was difficult to make it profitable in a low-margin industry,” said Taylor. “We use machine learning to find out the best approach for you. And then we put it to instant action, automatically.”